How do states determine which intimate relationships count as "family"? Why do some relationships receive legal entitlements - such as custody, inheritance, or domestic violence protection - while others, emotionally and materially similar, are excluded? This dissertation introduces the Intimacy Contract, a novel conceptual framework that illuminates how states selectively recognize and protect certain intimate relationships while dismissing others, based on an ideological vision of family.
I develop this framework through two empirical cases, both involving relationships that lack default legal recognition but where individuals nonetheless seek state rights and protections.
The first case focuses on women in extramarital relationships in India, a country with one of the highest reported rates of such relationships globally. Drawing on an original dataset of 2,721 appellate court cases (2000-2024), I show that despite anti-polygamy laws, Indian courts sometimes extend spousal entitlements to women in these relationships. Specifically, courts are 56 percent more likely to grant recognition when women are in relationships that resembles the state's ideal marriage - a conjugal life marked by shared household and children, a traditional gender dynamic where the woman is virtuous and vulnerable, and shared religiosity where the couple solemnize their relationship with a ritual and share the same faith. This recognition drops to 22 percent when these features are absent in the relationship.
The second case examines non-biological caregivers in U.S. child custody cases. Over 2.4 million children in the U.S. are raised by individuals who are not their legal or biological parents. I analyze how courts assess caregiving practices alongside biological, racial and gendered assumptions to decide when non-biological caregivers qualify as "parents."
Methodologically, I use a mixed-method approach that combines computational, quantitative and qualitative tools. I employ large language models (LLMs) to classify thousands of court judgments, identifying semantically nuanced case details that traditional pattern matching approaches tend to miss. These classifications are validated through a human-in-the-loop strategy involving iterative prompt refinement, random sampling, and coding safeguards. I then interpret the resulting patterns through statistical modeling, close readings of select cases, and interviews with legal practitioners and scholars.
By centering relationships as the unit of analysis, this project offers a novel account of state power that reveals its reach into the seemingly private realm of intimate relationships: the family. Focusing on the family reveals a system of power that is not reducible to gender, race, class or sexuality. Instead, it illuminates a system where some intimate relationships are protected over others through an ideological vision that defines what it means to belong in society.